{"id":"ee69c4dd-efca-4e9d-b74e-4ea732e234f9","arxiv_id":"2506.23359","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":8.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Every immersed 2-sphere with Willmore energy at most 12π admits an energy-nonincreasing regular homotopy to a round sphere or to a surface in the explicit family J, giving exactly four regular homotopy classes below 12π.","lead":"This paper maps the Willmore energy landscape of immersed spheres below energy 12π: every such surface deforms, without increasing energy, into one of four standard shape families. It also identifies exactly which starting surfaces force unavoidable singularities of the Willmore flow, and extends the Li-Yau inequality at 12π.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Four-class classification and 'unavoidable iff J' rest on imported companion-preprint results; if Theorem D, the F-values, or generic triple-point persistence fail, the lower-bound half collapses.","rationale":"The reader's weakest assumption correctly identifies the companion preprint [Sei25] as the most load-bearing input. I find no independent internal inconsistency in the gluing construction that produces the energy-decreasing homotopies of Theorem 1.2, and the 'at most four classes' part of Theorem 1.4 is plausible. However, the classification's lower bound and the unavoidable-singularity characterization cannot be certified from this manuscript alone: they rely on Theorem D, the F-invariant values, and generic triple-point persistence, all imported from a same-author companion preprint. Those results could be correct, but the present paper provides no verification, and the proof of Corollary 1.3 explicitly depends on the generic persistence in [Sei25]. The round-sphere edge case in Theorem 1.2 (where W(H_t)<4π is impossible) is a real but minor statement-level flaw that can be patched by excluding round spheres or allowing H_1=f_0; it does not affect the classification. Since the identified concern is exactly the one that makes the verdict conditional rather than accept, I recommend leaving the reader's CONDITIONAL verdict unchanged.","tokens_in":26854,"tokens_out":25054,"duration_ms":262018,"concrete_test":"Obtain [Sei25] and independently verify the three imported inputs: (i) confirm that Theorem 1.7 (every regular homotopy from j∈J to an embedded sphere has triple points) is proved without citing the present manuscript; (ii) recompute F(j) and F(±e) for the explicit rotationally symmetric model j of Figure 1 directly from the definition of F in [Sei25, Section 2], checking F(j)=e_{±3} and F(±e)=e_{±1}; (iii) verify that the triple point produced by Theorem 1.7 can be chosen generic, so that nearby immersions also have triple points, justifying the strict inequality W(H_t)>12π in Corollary 1.3. If any of these checks fails, the lower-bound half of Theorem 1.4 and Proposition 1.6 are not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The upper-bound direction of the main classification (at most four regular homotopy classes in I(ω)) is supported by the gluing construction in Section 2, and that part is internally plausible. The load-bearing weakness is the lower-bound direction: Theorem 1.4's assertion that j is not regularly homotopic to e or -e inside I(ω), Proposition 1.6's characterization of unavoidable singularities, and Proposition 1.9 all depend on Theorem D ([Sei25, Thm 1.7]), the invariant values F(±e)=e_{±1} and F(j)=e_{±3} ([Sei25, Ex 5.2]), and the generic-triple-point persistence used in Corollary 1.3 (Appendix A). None of these is proved or even summarized in the present manuscript; they are imported from a same-author preprint. If Theorem D's proof in [Sei25] relies on results from this manuscript (circular dependence), or if the computation F(j)=e_{±3} for the explicit rotationally symmetric model in Figure 1 is incorrect, then the lower bound 'at least four classes' fails, the infimum statement for J± is unsupported, and the 'only if' direction of Proposition 1.6 collapses. The present paper's construction cannot certify this lower bound on its own.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the sublevel sets I(omega) of the Willmore energy on the space of smooth immersions of S^2 into R^3. The main theorem (Theorem 1.2) asserts that every such immersion with energy at most 12 pi admits a regular homotopy of strictly decreasing energy whose terminal surface is either a round sphere or one of the model surfaces in the family J. Combining this with an invariant for triple-point-free immersed spheres imported from the companion preprint [Sei25], the paper derives a four-class description of I(omega) for omega in (8 pi, 12 pi], a classification of unavoidable singularities of the Willmore flow (Proposition 1.6), an extension of the Li-Yau inequality at 12 pi (Proposition 1.9), consequences for sphere eversions, and a partial classification below 16 pi. The constructive upper-bound part is carried out by gluing catenoid pieces to round spheres with explicit energy estimates, using the Lamm-Nguyen blow-up classification [LN15] and the Kuwert-Schaetzle convergence theorem [KS04].","tokens_in":27036,"tokens_out":17505,"duration_ms":178068,"significance":"If the results are correct, the paper gives a complete description of the regular homotopy classes of the Willmore energy landscape below 12 pi and identifies exactly which initial surfaces of energy at most 12 pi lead to unavoidable singularities. This would be a substantial advance in the global analysis of the Willmore functional. The paper's strength is its explicit gluing construction: the energy bookkeeping in Proposition 2.7, the estimates in Lemma 2.4, and the detailed computations in Lemma A.1 give a concrete, checkable pathway from a catenoid blow-up to either a round sphere or a surface in J. The upper-bound direction is internally plausible. However, the lower-bound direction and the classification of unavoidable singularities rest on Theorem D, the invariant values F(+-e)=e_{+-1}, F(j)=e_{+-3}, and a generic-triple-point persistence assertion, all imported from the same-author companion preprint [Sei25] without proof or summary in the present manuscript. In addition, the main theorem as stated is false for round initial data because the required strict energy decrease is impossible when W(f)=4 pi. These points make the paper conditional rather than self-contained.","major_comments":[{"comment":"Theorem 1.2 is false as stated for round initial data. If f is a round sphere, then W(f)=4 pi and every immersed sphere has Willmore energy at least 4 pi, with equality only for round spheres; hence no regular homotopy H with H_0=f can satisfy W(H_t)<W(f) for all t in (0,1]. The proof in Section 2.3 also implicitly assumes omega = W(f0)-W(f_{s0}) > 0, which fails for round f0. The theorem should be restricted to non-round initial data, or the strict inequality should be weakened in a way that is still sufficient for the later applications. Since the round spheres are already the endpoints e and -e in the classification, this is a fixable statement error, but it is load-bearing because Theorem 1.2 is cited as the main theorem.","section":"Section 1.3, Theorem 1.2"},{"comment":"The lower-bound half of the paper's central classification is not proved in this manuscript. The assertion that j in J is not regularly homotopic to e or -e inside I(omega), the 'only if' direction of Proposition 1.6, and Proposition 1.9 all depend on Theorem D ([Sei25, Theorem 1.7]), on the invariant values F(+-e)=e_{+-1} and F(j)=e_{+-3} ([Sei25, Example 5.2]), and on the claim that regular homotopies from J to embedded spheres have generic triple points. None of these results is stated in sufficient detail or proved here, and [Sei25] is a same-author companion preprint. If any of these imported results fails, the 'at least four classes' part of Theorem 1.4 collapses, and the characterization of unavoidable singularities is unsupported. The paper should either include proofs or sufficiently detailed statements of the imported results, or the main conclusions should be explicitly conditioned on the companion preprint.","section":"Section 1.3, Theorem 1.4; Section 1.4, Proposition 1.6; Section 1.5, Proposition 1.9"},{"comment":"The proof of Corollary 1.3 uses a 'generic triple-point persistence' assertion that is stronger than the quoted Theorem D. Theorem D only says that every regular homotopy joining some j in J to an embedded sphere has triple points; Corollary 1.3 needs the existence of a time at which the triple point is generic, so that all immersions in a C^infinity neighborhood also have triple points. This genericity statement is asserted to follow from [Sei25], but it is not formulated precisely nor proved. Since the strict inequality W(H_t)>12 pi is needed to separate the J-classes from the round classes in I(12 pi), this missing formulation is a load-bearing gap.","section":"Appendix A, Corollary 1.3"}],"minor_comments":[{"comment":"The abstract says a homotopy whose Willmore energy 'does not exceed' that of the initial surface, while Theorem 1.2 states the stronger strict inequality W(H_t)<W(f) for all t in (0,1]. These formulations should be reconciled, especially in light of the round-sphere issue raised above.","section":"Abstract and Section 1.3"},{"comment":"The definition says W_alpha(lambda,R,delta) is 'the Willmore energy of some catenoid sphere' in CatSph_alpha. Since the set CatSph_alpha may contain different parametrizations, it would be clearer to define W_alpha and W_beta as the energy of the specific rotationally symmetric gluing constructed in the preceding paragraph, and to note that this value is independent of the chosen parametrization.","section":"Section 2.2, Definition 2.3"},{"comment":"In the proof of Lemma 2.4(i), the constant delta_0 is set to 2 pi / C, but the statement requires delta_0 in (0,1/2). The definition should be delta_0 = min(1/2, 2 pi / C) to make the stated range valid.","section":"Section 2.2, proof of Lemma 2.4"},{"comment":"The text says that by translation and rotation of F_t one can ensure F_t(p)=0 and T_p F_t = R^2 x {0} for all t, but it does not explicitly justify that these ambient motions can be chosen continuously in t. This is likely true, but a sentence explaining the continuity would remove ambiguity.","section":"Section 2.3, Step 5.2 of Proposition 2.7"},{"comment":"There is a typo: 'The discussion aboe' should read 'The discussion above'.","section":"Section 2.4, Remark 2.11"}],"recommendation":"major_revision","confidential_remarks":"The main risk is the heavy reliance on the companion preprint [Sei25] for the lower-bound half of the classification. I would recommend that the editor require the authors to either include the relevant statements and proofs from [Sei25] in the paper or verify that the companion preprint has been independently accepted and posted in final form. The round-sphere counterexample to Theorem 1.2 is easy to fix, but it should be fixed before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, Theorem 1.2 as stated is false for round spheres: if W(f)=4π, the desired homotopy would have to lower energy below 4π, which Li-Yau forbids. That is an easy patch—exclude round spheres or allow equality—but it needs to be made. Second, the upper-bound direction (energy-decreasing homotopy to a round sphere or a model in J) is self-contained and looks solid; the lower-bound half of the four-class classification, the 'unavoidable iff J' characterization, and the Li-Yau-type extension all import Theorem D, the invariant values, and the triple-point persistence from the companion preprint [Sei25] by the same second author. If any of those fail, the classification collapses.\n\nWhat is genuinely new: the gluing construction in Section 2 with explicit energy bookkeeping, extending Kuwert-Schätzle's 8π theory to 12π. The strict energy chain W(H_t)<W(f0) is checked with slack ω/5 per gluing, and the catenoid-sphere models are concrete. The four regular homotopy classes in I(ω) for ω∈(8π,12π] and the classification of unavoidable singularities as exactly J are substantial steps in the Willmore landscape program.\n\nSoft spots: the round-sphere edge case makes Theorem 1.2, Lemma 2.8, and Proposition 2.7 overstate their validity. The proof of Corollary 1.3's strict inequality is abbreviated; it relies on generic triple-point persistence from [Sei25], and the no-critical-point-at-12π argument is sketched. The dependency on [Sei25] is the main risk: this manuscript cannot certify its own lower bound. I did not see circularity within the paper, but the referee must see [Sei25] and verify Theorem D and the F-values. Proposition 1.10's proof is independent and fine.\n\nWho this is for: anyone working on Willmore flow, energy landscapes, or sphere eversions. It deserves a serious referee, and the authors should be asked to fix the statement and to clarify/make available the companion's role.\n\nRecommendation: send to peer review conditional on access to [Sei25] and a corrected main theorem.","headline":"Genuine progress on the Willmore landscape below 12π, but the main theorem as stated fails for round spheres and the lower-bound half leans on a same-author companion preprint.","tokens_in":27804,"tokens_out":6029,"would_cite":true,"duration_ms":62159,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["53C42","53E40","57R42"],"pacs":[],"model":"deepseek-v4-flash","headline":"Every immersed 2-sphere with Willmore energy at most $12\\pi$ admits an energy-decreasing regular homotopy to either a round sphere or a surface in the model family $J$.","keywords":["Willmore energy","Willmore flow","regular homotopy class","immersed 2-spheres","catenoid blow-up","unavoidable singularity","Li-Yau inequality","triple-point-free invariant"],"falsifier":"Find a triple-point-free immersed sphere $f:S^2\\to\\mathbb{R}^3$ with $W(f)\\le 12\\pi$ whose invariant $F(f)$ takes a value outside $\\{e_{-3},e_{-1},e_1,e_3\\}$; Proposition 1.9 would then be false. Alternatively, construct a regular homotopy below $12\\pi$ from some $j\\in J$ to a round sphere, which would contradict Theorem D and collapse Theorem 1.4.","tokens_in":2232,"feed_emoji":"🔵","tokens_out":2146,"duration_ms":100413,"temperature":0.7,"pith_summary":"The paper proves that the low-energy part of the Willmore energy landscape for immersed 2-spheres in $\\mathbb{R}^3$ is completely described by four regular homotopy classes. More precisely, any smooth immersion $f:S^2\\to\\mathbb{R}^3$ with Willmore energy at most $12\\pi$ can be deformed, without ever raising its energy, to either a round sphere or one of the model surfaces in a family $J$. Combined with an invariant from a companion paper, this implies that the sublevel set $I(\\omega)$ has exactly four regular homotopy classes for $8\\pi<\\omega\\le 12\\pi$, and that the only initial surfaces whose Willmore flow singularities are unavoidable are precisely the surfaces in $J$. The result matters because it turns a global analytic question about the Willmore flow into a manageable topological classification, and it extends the Li–Yau inequality at the $12\\pi$ threshold to a large class of triple-point-free immersed spheres.","feed_headline":"Low-energy spheres deform to round spheres or one model family","feed_subtitle":"Below 12π every immersed sphere flows or deforms to a round sphere or to a model surface in the family J.","key_machinery":"The argument is carried by a gluing construction around the catenoid blow-up predicted by the blow-up classification for the Willmore flow below $16\\pi$: when $W(f_0)\\le 12\\pi$, the only possible blow-up at a singularity is a catenoid. The authors cut the surface along the catenoid neck, replace each side by a flat disk, and use the theorem that Willmore flow below $8\\pi$ converges to a round sphere. The two flowing hemispheres are glued back to the catenoid at every time, producing catenoid spheres of two types, $\\alpha$ and $\\beta$; type $\\beta$ can be shrunk to lower the energy below $8\\pi$ and then flowed to a round sphere, while type $\\alpha$ gives exactly the model family $J$. The companion invariant $F$ for triple-point-free immersed spheres is the topological tool that separates the four classes: it is constant along triple-point-free regular homotopies and takes values $e_{\\pm 1}$ on round spheres and $e_{\\pm 3}$ on $J$.","core_discovery":"The central claim, Theorem 1.2, is that energy below $12\\pi$ forces a very specific global shape: every smooth immersed sphere with $W(f)\\le 12\\pi$ admits a regular homotopy $H$ with $H_0=f$, $W(H_t)<W(f)$ for every positive time, and $H_1$ either a round sphere or an element of $J$. Theorem 1.4 then says that for $\\omega\\in(8\\pi,12\\pi]$ the sublevel set $I(\\omega)$ contains exactly four regular homotopy classes: two containing the round spheres of opposite orientation and two containing the model surfaces $\\pm j$, whose infimum energy is $8\\pi$ but not attained. Proposition 1.6 characterizes unavoidable singularities: a flow starting at $f_0$ with $W(f_0)\\le 12\\pi$ develops an unavoidable singularity exactly when $f_0\\in J$. Proposition 1.9 extends the Li–Yau inequality at $12\\pi$: a triple-point-free immersion with invariant $F(f)$ outside $\\{e_{\\pm 1},e_{\\pm 3}\\}$ must have $W(f)>12\\pi$.","pith_inferences":["If the companion invariant were replaced by a quadruple-point-free analogue, the same strategy would likely prove the $16\\pi$ conjecture once an explicit sphere eversion through a halfway model is shown to be free of quadruple points; the paper explicitly raises this as Question 1.12.","The four-class structure suggests a genuine topological phase change at $12\\pi$: numerical experiments could test whether energy-decreasing paths from random low-energy immersions land on round spheres or on the $J$-model family, and whether the infimum $8\\pi$ is approached by catenoid-sphere gluings.","The cut-and-glue procedure described as a possible Willmore flow with surgery is a testable algorithmic construction: implementing it at a catenoid blow-up and checking whether the resulting flow converges to one or two round spheres would provide a concrete dynamical realization of the classification.","For spheres of revolution, Proposition 1.10's lower bound $W(f)>4\\pi(|\\tau|+1/2)$ is sharp only in the limit; constructing explicit near-equality examples by gluing round spheres with catenoid necks would clarify how close the inequality is to being attained."],"forward_implications":["For every $\\omega\\in(8\\pi,12\\pi]$, the sublevel set $I(\\omega)$ splits into exactly four regular homotopy classes, so the energy landscape below $12\\pi$ is known at the level of regular homotopy classes.","If a smooth immersion with energy at most $12\\pi$ fails to converge to a round sphere under the Willmore flow, then the initial surface can be deformed, without energy increase, to a model surface in $J$; the singularity is unavoidable precisely for initial data in $J$.","The Li–Yau inequality is extended at the $12\\pi$ threshold: a triple-point-free immersed sphere with invariant outside $\\{e_{-3},e_{-1},e_1,e_3\\}$ has Willmore energy greater than $12\\pi$, so the new invariant $F$ controls the complexity of such surfaces.","For any initial surface with $W\\le 12\\pi$ that admits a triple-point-free regular homotopy to a round sphere, there is also an energy-nonincreasing such homotopy; this is the low-energy version of the statement needed for an optimal sphere eversion.","Below $16\\pi$, if no Enneper-type blow-up occurs, the same gluing method deforms the surface without energy increase to a round sphere or to an $\\alpha$-connected catenoid or trinoid model, giving a partial landscape description up to $16\\pi$."],"supporting_citations":[{"why":"Supplies the invariant $F$ for triple-point-free immersed spheres, its values $F(\\pm e)=e_{\\pm1}$ and $F(j)=e_{\\pm3}$, and Theorem D that every regular homotopy from $j\\in J$ to an embedded sphere has triple points; these are the lower-bound half of the classification.","marker":"[Sei25]"},{"why":"Classifies blow-up limits of the Willmore flow below $16\\pi$ and shows only catenoid blow-ups occur below $12\\pi$, the starting point of the gluing construction.","marker":"[LN15]"},{"why":"Proves long-time existence and convergence to a round sphere for Willmore flow of spheres with energy at most $8\\pi$, used after cutting and gluing the catenoid.","marker":"[KS04]"},{"why":"Proves every regular homotopy between oppositely oriented round spheres has a quadruple point, separating the two sphere classes from the two $J$-classes.","marker":"[MB81]"},{"why":"Provides the sharp lower bound $4\\pi n$ for immersions with an $n$-tuple point, used to get $W\\ge 12\\pi$ for triple points and $W\\ge 16\\pi$ for quadruple points.","marker":"[LY82]"},{"why":"Constructs surfaces of revolution in $J$ whose Willmore flow develops singularities, connecting the model family $J$ to actual flow behavior.","marker":"[Bla09]"},{"why":"Classifies smooth Willmore spheres and shows that below $16\\pi$ the only smooth Willmore spheres are round spheres, used to identify the endpoint of the flow.","marker":"[Bry84]"}],"fun_headline_variants":["Low-energy spheres deform to round spheres or one model family","Energy ≤12π spheres are round or in one model family","Willmore flow: low-energy spheres avoid singularities unless in J","Energy below 12π forces round sphere or model family"],"cache_read_input_tokens":29568,"weakest_assumption_plain":"The entire classification depends on the companion paper's invariant and on Theorem D asserting that every regular homotopy from a model surface $j\\in J$ to an embedded sphere must pass through at least one triple point; if that theorem were false, the four-class count and the 'unavoidable singularities exactly $J$' statement would both collapse.","fun_headline_variants_meta":{"raw":{"variants":["Low-energy spheres deform to round spheres or one model family","Energy ≤12π spheres are round or in one model family","Willmore flow: low-energy spheres avoid singularities unless in J","Energy below 12π forces round sphere or model family"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00045,"raw_usage":{"total_tokens":2275,"prompt_tokens":961,"completion_tokens":1314,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":577,"completion_tokens_details":{"reasoning_tokens":1244}},"tokens_in":577,"tokens_out":1314,"duration_ms":12070,"temperature":1.0,"reasoning_tokens":1244,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T21:47:55.675834+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Find a triple-point-free immersed sphere $f:S^2\\to\\mathbb{R}^3$ with $W(f)\\le 12\\pi$ whose invariant $F(f)$ takes a value outside $\\{e_{-3},e_{-1},e_1,e_3\\}$; Proposition 1.9 would then be false. Alternatively, construct a regular homotopy below $12\\pi$ from some $j\\in J$ to a round sphere, which would contradict Theorem D and collapse Theorem 1.4.","supporting_citations":[],"review_version":1}